Epilepsy is a common disorder. At present there is insufficient data on the natural history and prognosis of convulsive disorders. The present research plan is designed to prospectively investigate the risks of seizure recurrence in twp settings: 1) Following a single unprovoked seizure 2) Following discontinuation of anticonvulsant medications in patients with two or more seizures who have been seizure free for two or more years. In the case of first seizure, based on prior studies, two principal hypotheses are proposed. First, that the majority of patients with a single unprovoked seizure will not have further seizures and thus do not require anticonvulsant therapy; second, that there are predictive variables such as the EEG and neurological examination which will allow selection of a subgroup with a high recurrance risk. Children and adolescents with a first provoked seizure will be evaluated with a detailed history, neurological examination and an EEG. With few exceptions outlined in the research plan they will not be treated with long term anticonvulsant therapy. They will then be followed for several years. Those with a second seizure will be treated. The data will be analyzed with special emphasis on the probability of recurrence and on identifying significant prognostic variables. Based on prior studies, three hypotheses are advanced regarding children with epilepsy who have been seizure free for two or more years: first, that over 70% will remain seizure free when medications are discontinued; second, that utilizing a few predictive variables, the EEG, age of onset and seizure type,one can identify subgroups with excellent and poor prognosis; third, that most children who will relapse will do so within two years. Patients who are seizure free for two or more years will be evaluated and medications discontinued (see research plan for exceptions). They will then be prospectively followed for seizure recurrence. Data will be analyzed in terms of recurrence rate, prognostic variables and time to recurrence. An attempt will be made to validate the importance of the predictive variables which were found to be predictive in previous smaller scale trials. The data obtained from this research will be important in deciding which patients with a first seizure need to be treated and how long one needs to treat patients who are well controlled. Both these issues are of prime importance to the clinician dealing with epilepy.